top of page

From Storytelling to Compliance: Reimagining Risk Management With Reflect AI


From Storytelling to Compliance: Reimagining Risk Management With Reflect AI - Julian Guthrie

As an award-winning journalist and a New York Times best-selling author, my career has revolved around the power of words and their profound impact on society. The transition from journalism and nonfiction books to AI, by founding and leading a tech startup called Alphy, is a dramatic story unto itself — one I expect to write a book about in the coming years — and driven by a deep love of language and a passion for new technology. 


My journey began with a relentless pursuit of truth and a commitment to telling stories that matter. As a journalist, I delved into diverse subjects, from technology and entrepreneurship to human rights and social justice. These experiences honed my ability to dissect complex issues, communicate effectively, and uncover the underlying narratives that shape our world. This skill set has been invaluable in my role as the founder and CEO of Alphy, where we tackle one of the most critical challenges facing companies today: how people communicate with one another.


I also learned during my two decades in journalism that there are always multiple sides to a story and that it's important to meet people where they’re at — and to listen. Listen to what they say, the words they use, and the tone, and understand that our expressions reflect our experiences. Writing books about some of the world’s most successful people, from Larry Ellison and Peter Diamandis to the women of Alpha Girls, reinforced my belief that words give us the power to advance or retreat, do good or do harm. 


Addressing Communication


At Alphy, where we have built a proprietary AI classifier called Reflect AI to flag and detect harmful, unlawful, and unethical language in digital communication, we are trying to look anew at how companies nurture and manage internal codes of conduct and external regulatory oversight. We are reimagining solutions that can meet the complex demands of modern enterprise companies, beginning with those in financial services. We are asking experienced heads of surveillance, compliance, and conduct to expand how they define, monitor, remediate, and prevent lawsuits, reputational damage, and high turnover that stems from unlawful and unethical communication at work.


Traditional compliance and risk management solutions have long been focused on identifying rigidly defined issues, whether money laundering, antitrust violations, and, more recently, how employees are moving conversations to off-channel communications. While these are important and high-stakes, they represent the tip of the iceberg. At Alphy, we believe that compliance must encompass the full spectrum of human communication with its gradations of harm, from anger and racism or sexism to threats of bodily harm. This is where our groundbreaking Reflect AI technology comes into play.


Real-world Impact


Here are two stark examples where Reflect AI could have saved institutions tens of millions of dollars and/or reputational costs: 


Citigroup’s fairness in lending problem. The banking giant was fined $24.5 million in 2023 for blocking Armenian Americans from obtaining credit cards. The company used the discriminatory assumption that applicants with last names ending in -ian or -yan were more likely to be involved in the so-called “Southern California Armenian Mafia” and to commit crime and fraud. Reflect AI can detect racism based on ethnicity or country of origin. 


The FDIC’s disastrous culture of harassment. The Federal Deposit Insurance Corporation (FDIC) is under scrutiny by Congress after an explosive Wall Street Journal report and an independent consultant’s inquiry showed the agency is rife with sexual harassment and misogyny, leading women to abandon their jobs as regulators. Imagine being told by your manager, “Obviously, if I walked into this room and you were naked, I’d f— you right now,” or “Women need to use sex to get ahead at the FDIC,” as female employees alleged. Moreover, a dozen human resource officers have been the subject of employee complaints for their discriminatory language in recent years, one of the complaints reportedly leading to a $100,000 settlement, the Journal reported. Reflect AI can also detect gender discrimination and sexual harassment, as well as 24 other types of language.


Reflect AI offers real-time feedback to employees about the potential risks in their communications. In February, in a federal disability discrimination case, a New York jury awarded a woman $1.675 million after a distribution company declined to interview her upon learning she was deaf and for refusing to hire her for two entry-level jobs for which she applied. Imagine receiving a gentle nudge alerting you that a particular phrase might come across as racist or that a comment could be perceived as ableist?



“No credit card application approvals for Armenians.”

Racial discrimination prompt from Reflect AI





“You have been in this country long enough, you should adapt to the way our country runs!”

Racist prompt from Reflect AI





“I don't think having a wheelchair user as our front desk receptionist is a good look for this company.” 

Disabilities discrimination prompt from Reflect AI





“He’s deaf, he can’t flip burgers at Wendy’s — how we would talk to him?”

Ableist prompt from Reflect AI






This proactive approach not only helps individuals become more aware of their communication but also mitigates harm before it escalates into serious issues for the team and the company.


The history of compliance solutions in financial services is a testament to the evolving challenges and the need for continuous innovation. Compliance has always been a manual process, relying heavily on human oversight. With the advent of digital technologies, automated systems began to emerge, offering more efficient ways to monitor and enforce regulations. And with AI and machine learning, systems to detect harmful keywords or patterns of words began to take hold. 


A New Approach


However, these systems fell short of addressing the nuanced and context-dependent nature of human communication — until now. We are at a pivotal moment in which artificial intelligence has the potential to revolutionize compliance technology and Reflect AI by Alphy is at the forefront of this promise.


Our AI training team at Alphy brings a unique blend of linguistics and journalism expertise. We have trained our AI on trillions of human utterances, enabling it to grasp the subtleties and nuances that traditional systems overlook. This expertise allows us to provide more accurate feedback, helping employees navigate the complexities of modern communication with greater awareness while providing strong monitoring for companies.


We spend time as a team discussing things like the differences between timid, self-deprecating, and diminishing. Detecting these types of communication styles is important for feedback that steers employees to convey more authority. We all know someone who unconsciously undermines their authority by unnecessarily apologizing, being timid, or putting themselves down. It’s the difference between “Sorry I’m so late with this report,” and “Following up as requested,” or “Just wanted to say” and “Here’s what I think” or “I’m the worst in my department, I could never do that job” and “I’d need to brush up on my skills but sure, I’ll give it a shot.” We also have heated discussions around the difference between threats and retaliation and threats versus hyperbole.  


“You’ll be sorry you ever crossed me, I’ll make you pay…”

Threat detected prompt from Reflect AI





“We’re not going to promote him after he cooperated with HR on that EEOC investigation.”

Retaliation prompt from Reflect AI





“I’m going to kill him if he doesn’t finish that report by the deadline!”

Angry prompt from Reflect AI






Ongoing Innovation


At Alphy, we’re working on crafting a new vocabulary for compliance, one that speaks directly to the nuances of human interaction and experiences. We’ve combined our expertise in words with the transformative power of AI to deliver a new way of thinking about communication and compliance. We’d love to share what we’re doing with Reflect AI and see how it may help your team or company. You can reach us at sales@alphyco.com.



Julian Guthrie is the CEO and Founder of Alphy. 


Reflect AI by Alphy is an AI communication compliance solution that detects and flags language that is harmful, unlawful, and unethical in digital communication. Alphy was founded to reduce the risk of litigation from harmful and discriminatory communication while helping employees communicate more effectively.

Comments


bottom of page